https://github.com/spring-projects/spring-framework: Spring Object/Relational Mapping (Spring IO)
  • Juergen Hoeller 
/*
 * Copyright 2002-2017 the original author or authors.
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *      https://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package org.springframework.orm.jpa;

import java.sql.SQLException;

import javax.persistence.EntityManager;
import javax.persistence.PersistenceException;

import org.springframework.dao.support.PersistenceExceptionTranslator;
import org.springframework.jdbc.datasource.ConnectionHandle;
import org.springframework.lang.Nullable;
import org.springframework.transaction.TransactionDefinition;
import org.springframework.transaction.TransactionException;


SPI strategy that encapsulates certain functionality that standard JPA 2.1 does
not offer, such as access to the underlying JDBC Connection. This strategy is
mainly intended for standalone usage of a JPA provider; most of its functionality
is not relevant when running with JTA transactions.

In general, it is recommended to derive from DefaultJpaDialect instead of implementing this interface directly. This allows for inheriting common behavior (present and future) from DefaultJpaDialect, only overriding specific hooks to plug in concrete vendor-specific behavior. 
Author:Juergen Hoeller, Rod Johnson
See Also:
Since:2.0
/**
 * SPI strategy that encapsulates certain functionality that standard JPA 2.1 does
 * not offer, such as access to the underlying JDBC Connection. This strategy is
 * mainly intended for standalone usage of a JPA provider; most of its functionality
 * is not relevant when running with JTA transactions.
 *
 * <p>In general, it is recommended to derive from {@link DefaultJpaDialect} instead
 * of implementing this interface directly. This allows for inheriting common behavior
 * (present and future) from DefaultJpaDialect, only overriding specific hooks to
 * plug in concrete vendor-specific behavior.
 *
 * @author Juergen Hoeller
 * @author Rod Johnson
 * @since 2.0
 * @see DefaultJpaDialect
 * @see JpaTransactionManager#setJpaDialect
 * @see JpaVendorAdapter#getJpaDialect()
 * @see AbstractEntityManagerFactoryBean#setJpaDialect
 * @see AbstractEntityManagerFactoryBean#setJpaVendorAdapter
 */
public interface JpaDialect extends PersistenceExceptionTranslator {

	
Begin the given JPA transaction, applying the semantics specified by the
given Spring transaction definition (in particular, an isolation level
and a timeout). Called by JpaTransactionManager on transaction begin.

An implementation can configure the JPA Transaction object and then invoke begin, or invoke a special begin method that takes, for example, an isolation level. 
An implementation can apply the read-only flag as flush mode. In that case, a transaction data object can be returned that holds the previous flush mode (and possibly other data), to be reset in cleanupTransaction. It may also apply the read-only flag and isolation level to the underlying JDBC Connection before beginning the transaction. 
Implementations can also use the Spring transaction name, as exposed by the
passed-in TransactionDefinition, to optimize for specific data access use cases
(effectively using the current transaction name as use case identifier).

This method also allows for exposing savepoint capabilities if supported by the persistence provider, through returning an Object that implements Spring's SavepointManager interface. JpaTransactionManager will use this capability if needed. 
Params:
  • entityManager – the EntityManager to begin a JPA transaction on
  • definition – the Spring transaction definition that defines semantics
Throws:
See Also:
Returns:an arbitrary object that holds transaction data, if any (to be passed into cleanupTransaction). May implement the SavepointManager interface.
/**
	 * Begin the given JPA transaction, applying the semantics specified by the
	 * given Spring transaction definition (in particular, an isolation level
	 * and a timeout). Called by JpaTransactionManager on transaction begin.
	 * <p>An implementation can configure the JPA Transaction object and then
	 * invoke {@code begin}, or invoke a special begin method that takes,
	 * for example, an isolation level.
	 * <p>An implementation can apply the read-only flag as flush mode. In that case,
	 * a transaction data object can be returned that holds the previous flush mode
	 * (and possibly other data), to be reset in {@code cleanupTransaction}.
	 * It may also apply the read-only flag and isolation level to the underlying
	 * JDBC Connection before beginning the transaction.
	 * <p>Implementations can also use the Spring transaction name, as exposed by the
	 * passed-in TransactionDefinition, to optimize for specific data access use cases
	 * (effectively using the current transaction name as use case identifier).
	 * <p>This method also allows for exposing savepoint capabilities if supported by
	 * the persistence provider, through returning an Object that implements Spring's
	 * {@link org.springframework.transaction.SavepointManager} interface.
	 * {@link JpaTransactionManager} will use this capability if needed.
	 * @param entityManager the EntityManager to begin a JPA transaction on
	 * @param definition the Spring transaction definition that defines semantics
	 * @return an arbitrary object that holds transaction data, if any
	 * (to be passed into {@link #cleanupTransaction}). May implement the
	 * {@link org.springframework.transaction.SavepointManager} interface.
	 * @throws javax.persistence.PersistenceException if thrown by JPA methods
	 * @throws java.sql.SQLException if thrown by JDBC methods
	 * @throws org.springframework.transaction.TransactionException in case of invalid arguments
	 * @see #cleanupTransaction
	 * @see javax.persistence.EntityTransaction#begin
	 * @see org.springframework.jdbc.datasource.DataSourceUtils#prepareConnectionForTransaction
	 */
	@Nullable
	Object beginTransaction(EntityManager entityManager, TransactionDefinition definition)
			throws PersistenceException, SQLException, TransactionException;

	
Prepare a JPA transaction, applying the specified semantics. Called by
EntityManagerFactoryUtils when enlisting an EntityManager in a JTA transaction
or a locally joined transaction (e.g. after upgrading an unsynchronized
EntityManager to a synchronized one).

An implementation can apply the read-only flag as flush mode. In that case, a transaction data object can be returned that holds the previous flush mode (and possibly other data), to be reset in cleanupTransaction. 
Implementations can also use the Spring transaction name to optimize for
specific data access use cases (effectively using the current transaction
name as use case identifier).

Params:
  • entityManager – the EntityManager to begin a JPA transaction on
  • readOnly – whether the transaction is supposed to be read-only
  • name – the name of the transaction (if any)
Throws:
  • PersistenceException – if thrown by JPA methods
See Also:
Returns:an arbitrary object that holds transaction data, if any
(to be passed into cleanupTransaction)
/**
	 * Prepare a JPA transaction, applying the specified semantics. Called by
	 * EntityManagerFactoryUtils when enlisting an EntityManager in a JTA transaction
	 * or a locally joined transaction (e.g. after upgrading an unsynchronized
	 * EntityManager to a synchronized one).
	 * <p>An implementation can apply the read-only flag as flush mode. In that case,
	 * a transaction data object can be returned that holds the previous flush mode
	 * (and possibly other data), to be reset in {@code cleanupTransaction}.
	 * <p>Implementations can also use the Spring transaction name to optimize for
	 * specific data access use cases (effectively using the current transaction
	 * name as use case identifier).
	 * @param entityManager the EntityManager to begin a JPA transaction on
	 * @param readOnly whether the transaction is supposed to be read-only
	 * @param name the name of the transaction (if any)
	 * @return an arbitrary object that holds transaction data, if any
	 * (to be passed into cleanupTransaction)
	 * @throws javax.persistence.PersistenceException if thrown by JPA methods
	 * @see #cleanupTransaction
	 */
	@Nullable
	Object prepareTransaction(EntityManager entityManager, boolean readOnly, @Nullable String name)
			throws PersistenceException;

	
Clean up the transaction via the given transaction data. Called by
JpaTransactionManager and EntityManagerFactoryUtils on transaction cleanup.

An implementation can, for example, reset read-only flag and
isolation level of the underlying JDBC Connection. Furthermore,
an exposed data access use case can be reset here.

Params:
  • transactionData – arbitrary object that holds transaction data, if any
(as returned by beginTransaction or prepareTransaction)
See Also:
/**
	 * Clean up the transaction via the given transaction data. Called by
	 * JpaTransactionManager and EntityManagerFactoryUtils on transaction cleanup.
	 * <p>An implementation can, for example, reset read-only flag and
	 * isolation level of the underlying JDBC Connection. Furthermore,
	 * an exposed data access use case can be reset here.
	 * @param transactionData arbitrary object that holds transaction data, if any
	 * (as returned by beginTransaction or prepareTransaction)
	 * @see #beginTransaction
	 * @see org.springframework.jdbc.datasource.DataSourceUtils#resetConnectionAfterTransaction
	 */
	void cleanupTransaction(@Nullable Object transactionData);

	
Retrieve the JDBC Connection that the given JPA EntityManager uses underneath, if accessing a relational database. This method will just get invoked if actually needing access to the underlying JDBC Connection, usually within an active JPA transaction (for example, by JpaTransactionManager). The returned handle will be passed into the releaseJdbcConnection method when not needed anymore. 
This strategy is necessary as JPA does not provide a standard way to retrieve
the underlying JDBC Connection (due to the fact that a JPA implementation might not
work with a relational database at all).

Implementations are encouraged to return an unwrapped Connection object, i.e.
the Connection as they got it from the connection pool. This makes it easier for
application code to get at the underlying native JDBC Connection, like an
OracleConnection, which is sometimes necessary for LOB handling etc. We assume
that calling code knows how to properly handle the returned Connection object.

In a simple case where the returned Connection will be auto-closed with the EntityManager or can be released via the Connection object itself, an implementation can return a SimpleConnectionHandle that just contains the Connection. If some other object is needed in releaseJdbcConnection, an implementation should use a special handle that references that other object. 
Params:
  • entityManager – the current JPA EntityManager
  • readOnly – whether the Connection is only needed for read-only purposes
Throws:
  • PersistenceException – if thrown by JPA methods
  • SQLException – if thrown by JDBC methods
See Also:
Returns:a handle for the Connection, to be passed into releaseJdbcConnection, or null if no JDBC Connection can be retrieved
/**
	 * Retrieve the JDBC Connection that the given JPA EntityManager uses underneath,
	 * if accessing a relational database. This method will just get invoked if actually
	 * needing access to the underlying JDBC Connection, usually within an active JPA
	 * transaction (for example, by JpaTransactionManager). The returned handle will
	 * be passed into the {@code releaseJdbcConnection} method when not needed anymore.
	 * <p>This strategy is necessary as JPA does not provide a standard way to retrieve
	 * the underlying JDBC Connection (due to the fact that a JPA implementation might not
	 * work with a relational database at all).
	 * <p>Implementations are encouraged to return an unwrapped Connection object, i.e.
	 * the Connection as they got it from the connection pool. This makes it easier for
	 * application code to get at the underlying native JDBC Connection, like an
	 * OracleConnection, which is sometimes necessary for LOB handling etc. We assume
	 * that calling code knows how to properly handle the returned Connection object.
	 * <p>In a simple case where the returned Connection will be auto-closed with the
	 * EntityManager or can be released via the Connection object itself, an
	 * implementation can return a SimpleConnectionHandle that just contains the
	 * Connection. If some other object is needed in {@code releaseJdbcConnection},
	 * an implementation should use a special handle that references that other object.
	 * @param entityManager the current JPA EntityManager
	 * @param readOnly whether the Connection is only needed for read-only purposes
	 * @return a handle for the Connection, to be passed into {@code releaseJdbcConnection},
	 * or {@code null} if no JDBC Connection can be retrieved
	 * @throws javax.persistence.PersistenceException if thrown by JPA methods
	 * @throws java.sql.SQLException if thrown by JDBC methods
	 * @see #releaseJdbcConnection
	 * @see org.springframework.jdbc.datasource.ConnectionHandle#getConnection
	 * @see org.springframework.jdbc.datasource.SimpleConnectionHandle
	 * @see JpaTransactionManager#setDataSource
	 */
	@Nullable
	ConnectionHandle getJdbcConnection(EntityManager entityManager, boolean readOnly)
			throws PersistenceException, SQLException;

	
Release the given JDBC Connection, which has originally been retrieved via getJdbcConnection. This should be invoked in any case, to allow for proper release of the retrieved Connection handle. 
An implementation might simply do nothing, if the Connection returned by getJdbcConnection will be implicitly closed when the JPA transaction completes or when the EntityManager is closed. 
Params:
  • conHandle – the JDBC Connection handle to release
  • entityManager – the current JPA EntityManager
Throws:
  • PersistenceException – if thrown by JPA methods
  • SQLException – if thrown by JDBC methods
See Also:
  • getJdbcConnection
/**
	 * Release the given JDBC Connection, which has originally been retrieved
	 * via {@code getJdbcConnection}. This should be invoked in any case,
	 * to allow for proper release of the retrieved Connection handle.
	 * <p>An implementation might simply do nothing, if the Connection returned
	 * by {@code getJdbcConnection} will be implicitly closed when the JPA
	 * transaction completes or when the EntityManager is closed.
	 * @param conHandle the JDBC Connection handle to release
	 * @param entityManager the current JPA EntityManager
	 * @throws javax.persistence.PersistenceException if thrown by JPA methods
	 * @throws java.sql.SQLException if thrown by JDBC methods
	 * @see #getJdbcConnection
	 */
	void releaseJdbcConnection(ConnectionHandle conHandle, EntityManager entityManager)
			throws PersistenceException, SQLException;

}